Machine Learning Approach for Energy Consumption Prediction in Datacenters

Abstract

Cloud Computing represents the ideal solution for end-users either small medium enterprises or simple clients. This solution is given as a for clients to go from classic service concept to oriented service concept. Moreover, this paradigm collects a set of operations which made them a complex task to the managers. Since the coming of the Cloud Computing encourages service providers to deploy their services. These enormous services need some infrastructure services that are located in datacenters in order to execute them. Due to this use, Cloud infrastructure owners are concerned by the huge energy consumed during this execution. This problematic will affect the use of costs for the services providers. To tackle this problem, in this work, we present several models presented in machine learning methods in order to predict the energy to be consumed for the next use. These forecasts could help the infrastructure providers to propose a plan and some analytics to eliminate the waste of used resources during the execution of services. The implementation of this model has been provided in order to evaluate our system. The obtained results demonstrate the effectiveness of our proposed system.